CN113656603A

CN113656603A - Method and device for obtaining field description information

Info

Publication number: CN113656603A
Application number: CN202111032146.0A
Authority: CN
Inventors: 龚厚瑜
Original assignee: Beijing IQIYI Science and Technology Co Ltd
Current assignee: Beijing IQIYI Science and Technology Co Ltd
Priority date: 2021-09-03
Filing date: 2021-09-03
Publication date: 2021-11-16
Anticipated expiration: 2041-09-03
Also published as: CN113656603B

Abstract

The embodiment of the invention provides a field description information obtaining method and a field description information obtaining device, which relate to the technical field of data processing, wherein the field description information obtaining method comprises the following steps: acquiring user search behaviors aiming at video content, and extracting a first field for searching from the user search behaviors; determining a second field directly associated with the first field based on historical search behavior for the video content; searching a first entity which is directly associated with the first field from the constructed knowledge graph; according to the first entity, searching a second entity which indirectly has an association relation with the first field from the knowledge graph, and determining a third field which indirectly has an association relation with the first field based on the first association relation between the second field and the known field; and obtaining the description information of the first field according to the target description information. By applying the scheme provided by the embodiment of the invention, the accuracy of the obtained description information of the first field can be improved.

Description

Method and device for obtaining field description information

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a field description information obtaining method and apparatus.

Background

In a user search scenario, in a case where a user needs to search for video content, the user generally enters a field to be queried in a search box, so that a server can provide the user with content based on the field to be queried. In addition, in order to attract the user to search other contents provided by the server, the associated field of the field to be inquired can be recommended for the user after the user inputs the field to be inquired. For example, in the case where the user inputs the field to be queried, namely "soldier assault", the associated field may be the name of the lead actor of "soldier assault", such as "king treasure strong", the name of other tv shows with the same theme as "soldier assault", such as "bright sword", and the like.

In the prior art, other fields similar to the field to be queried can be generally determined as the associated fields based on the field description information. For one-time user search behavior of a user, a first field used for searching can be extracted from the user search behavior, historical search behavior of searching other fields except the first field used by each user within a preset time period is determined from the historical search behavior, and a second field used for searching is extracted from the determined historical search behavior. Theoretically, the degree of association of the search behavior performed by the user in a short time may be high, and the degree of association between the second field and the first field may be high, so that the description information of the first field may be obtained according to the description information of the second field.

However, in practical situations, fields used by a user for searching video content tend to be more divergent in content, that is, the degree of association between a first field and a second field may be lower, and the noise of the fields used for searching is higher, for example, if the user searches both an animation and an art program within a time period of a preset time duration, the first field is a field for searching the animation, the second field is a field for searching the art program, and the degree of association between the first field and the second field is lower. Thus, the accuracy of the description information of the determined first field may be low, subject to the influence of the second field having a lower degree of association.

Disclosure of Invention

The embodiment of the invention aims to provide a field description information obtaining method and device so as to improve the accuracy of the obtained field description information. The specific technical scheme is as follows:

in a first aspect of the embodiments of the present invention, a method for obtaining field description information is provided first, where the method includes:

acquiring user search behaviors aiming at video content, and extracting a first field for searching from the user search behaviors;

determining a second field directly having an association relation with the first field based on historical search behaviors aiming at the video content, wherein a time difference between a first search time and a second search time is less than or equal to a preset time difference, and the first search time is as follows: the time for searching by using the first field in the historical searching action is as follows: time to search using the second field in the historical search behavior;

searching a constructed knowledge graph for a first entity directly associated with the first field, wherein the knowledge graph is: composed of entities associated with video content;

according to the first entity, searching a second entity indirectly having an association relation with the first field from the knowledge graph, and determining a third field indirectly having an association relation with the first field based on the first association relation between the second field and the known field;

obtaining description information of the first field according to target description information, wherein the target description information comprises: the description information of the second field, the description information of the third field, the description information of the first entity, and the description information of the second entity.

In a second aspect of the embodiments of the present invention, there is also provided a field description information obtaining apparatus, including:

the field extraction module is used for acquiring user search behaviors aiming at video contents and extracting a first field for searching from the user search behaviors;

a field determination module, configured to determine, based on historical search behavior for video content, a second field directly associated with a first field, where a time difference between a first search time and a second search time is less than or equal to a preset time difference, where the first search time is: the time for searching by using the first field in the historical searching action is as follows: time to search using the second field in the historical search behavior;

an entity searching module, configured to search a constructed knowledge graph for a first entity directly associated with the first field, where the knowledge graph is: composed of entities associated with video content;

the entity field searching module is used for searching a second entity indirectly having an association relation with the first field from the knowledge graph according to the first entity, and determining a third field indirectly having an association relation with the first field based on the first association relation between the second field and the known field;

an information obtaining module, configured to obtain description information of the first field according to target description information, where the target description information includes: the description information of the second field, the description information of the third field, the description information of the first entity, and the description information of the second entity.

In a third aspect of the embodiments of the present invention, an electronic device is provided, which includes a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;

a memory for storing a computer program;

a processor for implementing the method steps of any of the first aspect when executing a program stored in the memory.

In a further aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, in which a computer program is stored, and the computer program, when executed by a processor, implements the method steps of any one of the above first aspects.

In a further aspect of embodiments of the present invention, there is also provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the method steps of any of the first aspects described above.

The field description information obtaining method provided by the embodiment of the invention obtains the user searching behavior aiming at the video content, and extracts the first field for searching from the user searching behavior; determining a second field with a time difference smaller than a preset time difference between the search time and the search time of the first field based on historical search behaviors aiming at the video content, and using the second field as a field directly associated with the first field; searching a constructed knowledge graph related to video content for a first entity directly associated with the first field; and according to the first entity, searching a second entity which indirectly has an association relation with the first field from the knowledge graph, and determining a third field which indirectly has an association relation with the first field based on the second field and the known first association relation between the fields. And obtaining the description information of the first field according to the description information of the second field, the description information of the third field, the description information of the first entity and the description information of the second entity.

As can be seen from the above, in addition to determining the second field directly associated with the first field based on the historical search behavior, the solution provided by the embodiment of the present invention also determines the first entity in the knowledge graph directly associated with the first field, and also determines the third field indirectly associated with the first field and the second entity. Therefore, by introducing the constructed knowledge graph, the first field not only has an association relationship with the known field, but also has an association relationship with the entity, and the association relationship with the first field is rich. Therefore, the nodes such as the fields and the entities having the association relation with the first field are abundant, and the description information of the first field obtained according to the description information of the abundant nodes is accurate.

In addition, for the first field with a lower occurrence probability in the history search behavior, it can be determined that there are fewer fields having an association relationship with the first field, and the accuracy of the description information of the first field obtained based on the description information of the smaller number of fields having an association relationship is lower. The embodiment of the invention introduces the constructed knowledge graph, and can search the entity having the association relation with the first field from the knowledge graph besides the field having the association relation with the first field, so that the number of the searched nodes having the association relation with the first field is larger, and the accuracy of the description information of the first field obtained based on the description information of the nodes with larger number is higher.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.

Fig. 1 is a schematic flowchart of a first field description information obtaining method provided in an embodiment of the present invention;

FIG. 2 is a schematic diagram of a behavior map provided in an embodiment of the present invention;

FIG. 3 is a schematic illustration of a knowledge-graph provided in an embodiment of the present invention;

FIG. 4 is a schematic diagram of a behavior knowledge graph provided in an embodiment of the present invention;

fig. 5 is a flowchart illustrating a second field description information obtaining method according to an embodiment of the present invention;

fig. 6 is a flowchart illustrating a third field description information obtaining method provided in an embodiment of the present invention;

fig. 7 is a schematic structural diagram of a first field description information obtaining apparatus provided in an embodiment of the present invention;

fig. 8 is a schematic structural diagram of a second field description information obtaining apparatus provided in an embodiment of the present invention;

fig. 9 is a schematic structural diagram of a third field description information obtaining apparatus provided in the embodiment of the present invention;

fig. 10 is a schematic structural diagram of an electronic device provided in an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived from the embodiments given herein by one of ordinary skill in the art, are within the scope of the invention.

Since the accuracy of the description information of the first field obtained in the prior art is low. To solve the problem, embodiments of the present invention provide a field description information obtaining method and apparatus.

In an embodiment of the present invention, a method for obtaining field description information is provided, where the method includes:

determining a second field directly having an association relation with the first field based on historical search behaviors aiming at the video content, wherein a time difference between a first search time and a second search time is less than or equal to a preset time difference, and the first search time is as follows: the time for searching by using the first field in the history searching behavior is as follows: time to search using the second field in the historical search behavior;

searching a first entity directly associated with the first field from a constructed knowledge graph, wherein the knowledge graph is as follows: composed of entities associated with video content;

according to the first entity, searching a second entity which indirectly has an association relation with the first field from the knowledge graph, and determining a third field which indirectly has an association relation with the first field based on the first association relation between the second field and the known field;

obtaining description information of the first field according to target description information, wherein the target description information includes: the description information of the second field, the description information of the third field, the description information of the first entity, and the description information of the second entity.

A field description information obtaining method and device provided by the embodiments of the present invention are described below with specific embodiments.

Referring to fig. 1, a flowchart of a first field description information obtaining method according to an embodiment of the present invention is shown, where the method includes the following steps S101 to S105.

S101: user search behavior for video content is obtained, and a first field for searching is extracted from the user search behavior.

Specifically, after detecting the user search behavior of the user for the video content, the first field may be extracted from the field used in the user search behavior.

The first field may be a complete field used in a user search action, or may be a field composed of partial characters included in the complete field.

In an embodiment of the present invention, the complete field may be subjected to word segmentation, and a keyword therein is extracted as the first field. Specifically, the word segmentation processing method in the prior art may be adopted to perform word segmentation and combing on the complete field to obtain the first field.

S102: a second field directly associated with the first field is determined based on historical search behavior for the video content.

And the time difference between the first search time and the second search time is less than or equal to a preset time difference. The first search time is: the time of searching using the first field in the historical search behavior. The second search time is: the time in the historical search behavior to search using the second field.

In addition, in the historical search behavior of each user, the user may use the first field to search multiple times, and thus there may be multiple first search times, and different second fields may be determined based on the different first search times.

The association relationship between the first field and the second field may be a directed relationship or a non-directed relationship, the association relationship between the first field and the second field is different in weight, and a larger weight indicates a higher degree of association between the first field and the second field.

Wherein the directed relationship may represent a context between a time of searching using the first field and a time of searching using the second field in the historical search behavior. It may be considered that in the case where the first time precedes the second time, there is an association in which the first field points to the second field. Conversely, if the first time is later than the second time, there is an association of the second field to the first field.

Alternatively, it may be considered that, when the first time is later than the second time, there is an association in which the first field points to the second field. Conversely, if the first time precedes the second time, it may be assumed that there is an association of the second field to the first field.

While the undirected relationship does not represent a contextual relationship between the first time and the second time. There may be an undirected association between the first field and the second field whether the first time is prior or the second time is prior.

In addition, the association relationship between the first field and the second field may be stored in a form of a triple, where the triple includes three elements, which respectively represent the first field, the second field, and the association relationship. For example, the first element is the identifier of the first field, the second element is the identifier of the association, and the third element is the identifier of the second field. The identifier of the field may be a number of the field or the field itself, the identifier of the association may be a type of the association, for example, an identifier indicating that a first time corresponding to the first field is before a second time corresponding to the second field, an identifier indicating that the first time corresponding to the first field is after the second time corresponding to the second field, and the like, and the identifier of the association may also be a weight of the association.

Specifically, in the history search behaviors of each user, the history search behavior in which the time difference between the corresponding second time and the corresponding first time is smaller than the preset time difference may be determined, and a field may be extracted from the determined history search behavior as the second field. For example, the preset time difference may be 5 minutes. The manner of extracting fields from the history search behavior is similar to that in step S101, and is not described herein again.

Step S102 may be implemented by step a, and will not be described in detail here.

S103: and searching the first entity which is directly associated with the first field from the constructed knowledge graph.

Wherein, the knowledge graph comprises: consisting of entities associated with video content. For example, the entities associated with video content described above may include movies, television shows, actors, dramas, directors, photographers, awards, movie titles, and so forth.

Specifically, the knowledge graph includes nodes corresponding to the entities and edges between the nodes, and the presence of the edges between the nodes indicates that an association relationship exists between the entities corresponding to the nodes on both sides of the edges. The knowledge graph may further include weights of association relationships between entities, where the weights represent degrees of association between entities having association relationships, and the weights of association relationships between different entities may be the same or different. If the weight of the association relation is not recorded in the knowledge graph separately, the weight between the entities may be defaulted as a first preset weight.

In addition, the association relationship between the first field and the first entity may represent: the first field is the same as the object corresponding to the first entity, for example, the first field is "wangbaoqiang", and the first entity is an entity recording information of wangbaoqiang.

The association relationship between the first field and the first entity may also represent: the object corresponding to the first field includes an object corresponding to the first entity, or the object corresponding to the first entity includes an object corresponding to the first field, for example, the first field is "wangbaoqiang soldier assault", the first field corresponds to both the object wangbaoqiang and the object soldier assault, the first entity is an entity recording information of wangbaoqiang, or the first field is "movie by wujing director", and the first entity is an entity recording information of warwolf.

Furthermore, the weight of the association relationship between the first field and the first entity may be set to a second preset weight.

In an embodiment of the present invention, the first entity may be determined through step B and/or step C, and the embodiment of the present invention is not described in detail herein.

S104: and according to the first entity, searching a second entity which indirectly has an association relation with the first field from the knowledge graph, and determining a third field which indirectly has an association relation with the first field based on the first association relation between the second field and the known field.

Specifically, since the association relationship between the entities is originally recorded in the knowledge graph, the entity in the knowledge graph, which has an association relationship with the first entity directly and/or indirectly, may be determined as the second entity. In addition, known association relationship may exist between known fields, and a field having an association relationship directly and/or indirectly with the second field may be determined from the known fields based on the association relationship as the third field. Furthermore, there may also be an association between a known field and an entity in the knowledge-graph, so a known field directly and/or indirectly associated with the first entity and the second entity may be determined as a third field, and an entity directly and/or indirectly associated with the second field and the third field may be determined as a third entity.

The association relationship between the known fields may be determined in a manner similar to step S102, and the association relationship between the known fields and the entities in the knowledge graph may be determined in a manner similar to step S103, which is not described herein again.

In addition, the distances between the different third fields and the first field are different, that is, the minimum number of fields or nodes through which the association relationship indirectly exists between the different third fields and the first field passes is different. For example, if there is an association between the first field and the known field X directly, there is no association between the first field and the known field Y directly, but there is an association between the field X and the field Y directly, the field Y indirectly has an association with the first field through 1 field X, and the distance between the field Y and the first field is 1.

The distance between the different second entities and the first field is also different, i.e. the minimum number of fields or nodes through which the indirectly existing associations between the different second entities and the first field pass is different. For example, if there is an association between the first field and the known field X directly and there is no association with the entity Z directly, but there is an association between the field X and the known field Y directly and there is an association between the entity Z and the field Y directly, then the entity Z indirectly has an association with the first field through 2 fields of the field X and the field Y, and the distance between the entity Z and the first field is 2.

In theory, the field with the larger distance to the first field or the association degree between the entity and the first field is smaller, so the embodiment of the present invention may select, as the third field, the field with the distance to the first field smaller than the first preset distance, from the fields indirectly associated with the first field. And selecting an entity with a distance less than a second preset distance from the first field as a second entity from the entities indirectly associated with the first field, so as to ensure that the determined association degree between the third field and the first field is higher and ensure that the determined association degree between the second entity and the first field is higher.

The first preset distance and the second preset distance may be the same or different.

In another embodiment of the present invention, the step S104 can be realized through the steps S103A-S103B, which will not be described in detail herein.

S105: and obtaining the description information of the first field according to the target description information.

Wherein, the target description information includes: the description information of the second field, the description information of the third field, the description information of the first entity, and the description information of the second entity.

Specifically, the description information may be represented in the form of a feature vector, and an average value of the feature vectors of the second field, the third field, the first entity, and the second entity may be calculated as the description information of the first field.

In addition, the second field has direct association with the first field, the first entity has direct association with the first field, the third field has indirect association with the first field, and the second entity has indirect association with the first field. Therefore, for the first field, the association degree between the second field and the first entity is higher, and the association degree between the third field and the second entity is lower, so that the description information of the second field and the first entity has a larger influence on the description information of the first field, and the description information of the third field and the second entity has a smaller influence on the description information of the first field. And the distances between the different third fields and the second entity and the first field are different, and the larger the distance is, the smaller the influence on the description information of the first field is. Moreover, the weights of different association relations are different, and the first field has a larger influence on the description information of the first field when the weights of the second field, the third field and the association relation between the first entity and the second entity are larger.

Therefore, when the description information of the first field is calculated, different weight values can be used for different second fields, third fields, first entities and second entities. In the case that the description information is expressed in the form of a feature vector, the description information of the first field may be obtained by performing weighted calculation, weighted average calculation, and the like on the second field, the third field, the feature vectors of the first entity and the second entity.

Furthermore, the description information of the first field may be obtained by using a graph embedding method, which may be a node2vec method or other methods in the prior art, and this is not limited in the embodiment of the present invention.

In an embodiment of the present invention, the known association relationship between the fields may be represented in a form of a behavior graph, where the behavior graph includes nodes and edges between the nodes, each node corresponds to one known field, and the edges between the nodes represent that the association relationship exists between the fields corresponding to the nodes on both sides of the edge.

After determining a second field in the known field, which has an association relationship with the first field directly, the first field may be added to the behavior graph as a new node, and an edge may be added between the new node and a node corresponding to the second field to indicate that an association relationship exists between the first field and the second field.

Referring to fig. 2, a schematic diagram of a behavior map according to an embodiment of the present invention is provided.

Wherein each circle represents each node and each connecting line represents each edge in the behavior graph. The node 1 corresponds to a first field, the nodes 2 to 4 correspond to known fields, a connecting line between the node 1 and the node2 indicates that an association relationship exists between the first field and the field corresponding to the node2, and the field corresponding to the node2 is a second field. The connecting line between the node2 and the node 3 indicates that a first association relationship exists between the field corresponding to the node2 and the field corresponding to the node 3, and the connecting line between the node2 and the node 4 indicates that a first association relationship exists between the field corresponding to the node2 and the field corresponding to the node 4.

Because the

nodes

3, 4 and 1 are not directly connected with each other, but can be connected with the node 1 through the node2, the fields corresponding to the

nodes

3 and 4 indirectly have an association relationship with the first field corresponding to the node 1, and the field corresponding to the

nodes

3 and 4 is the third field.

In another embodiment of the present invention, the kind of relationship represented by the association relationship between different entities may be different, for example, the kind of relationship may be: synonymy relations having the same meaning are indicated by the entity information, and an inclusion relation in which one entity includes another entity, and the like. For example, the association between the entity tom gruus and the entity attoo may be a synonymy relationship, and the association between the entity ancient drama and the entity qinggong drama may be an inclusion relationship.

The association relationship between the entities may be a directed relationship or an undirected relationship. For example, if there is a synonymous relationship between the entity M and the entity N, the meaning of the entity information of the entity M is the same as that of the entity information of the entity N for the entity M, and the meaning of the entity information of the entity N is the same as that of the entity information of the entity M for the entity N, the synonymous relationship is a undirected relationship. If there is an inclusion relationship between the entity M and the entity N, it means that the entity M includes the entity N, but the entity N does not include the entity M, so the inclusion relationship is a directed relationship.

Referring to fig. 3, a schematic diagram of a knowledge graph according to an embodiment of the present invention is provided.

Each circle represents each node in the knowledge graph, nodes 5 to 9 correspond to entities a to e, each connection line represents each edge in the knowledge graph, a connection line between the node 5 and the node 6 represents that an association relationship exists between the entity a and the entity b, a connection line between the node 5 and the node 7 represents that an association relationship exists between the entity a and the entity c, a connection line between the node 6 and the node 8 represents that an association relationship exists between the entity b and the entity d, a connection line between the node 7 and the node 8 represents that an association relationship exists between the entity c and the entity d, and a connection line between the node 8 and the node 9 represents that an association relationship exists between the entity d and the entity e.

In one embodiment of the present invention, the above-described knowledge-graph may be constructed based on the following steps D1-D3.

Step D1: and acquiring entities related to the video content, and determining the association relation among the entities.

Specifically, the entities related to the video content and the association relationship between the entities may be locally stored, or may be acquired from a network.

Step D2: and generating the triples based on the association relation among the entities.

Wherein, the elements in the triples are respectively: entity identifications of two entities having an association relationship and a type of association relationship between the entities.

Specifically, the entity identifier may be a name, a number, and the like of the entity.

Step D3: and constructing the knowledge graph based on the generated triples.

Specifically, each triple record has an association relationship between a pair of entities, and the association relationships between each pair of entities are combined to obtain the knowledge graph.

The knowledge graph can be represented by a set of the triples, or each triplet is used as an entry in a knowledge graph table to represent the knowledge graph in the form of the knowledge graph table.

In another embodiment of the present invention, in the case that the first field is represented by a node in the behavior-graph, an edge may be added between the node corresponding to the first field and the node in the knowledge-graph corresponding to the first entity, so as to connect the behavior-graph and the knowledge-graph to form the behavior-knowledge-graph.

Referring to fig. 4, a schematic diagram of a behavior knowledge graph according to an embodiment of the present invention is provided, where the behavior knowledge graph shown in fig. 4 is obtained by connecting the behavior knowledge graph shown in fig. 2 with the behavior knowledge graph shown in fig. 3. In the figure, the behavior map is above the dotted line, and the knowledge map is below the dotted line. The connection between the node 1 and the node 5 indicates that the first field directly has an association relationship with the entity a, the connection between the node 1 and the node 6 indicates that the first field directly has an association relationship with the entity b, and the entity a and the entity b are the first entity.

In addition, the

nodes

7, 8, 9 and the node 1 do not have direct connection, but all have connection with the node 1 via other nodes, so that the entity c, the entity d and the entity e indirectly have association with the first field, and the entity c, the entity d and the entity e are second entities.

In one embodiment of the present invention, the step S102 can be implemented by the following step a.

Step A: and determining a field with the target frequency greater than the preset frequency as a second field based on the historical search behavior aiming at the video content.

Wherein, the target frequency is: the time difference between the time for searching by using the field in the historical searching action and the first searching time is less than the frequency of the preset time difference.

Specifically, for a field, if the time difference between the first search time when each user uses the first field to perform the search and the time when each user uses the field to perform the search is smaller than the preset time difference, and the target frequency when the time difference is smaller than the preset time difference is greater than the preset frequency, each user may be considered to frequently search the first field and the field together in a short time, and the association degree between the first field and the field may be considered to be higher, and the field is determined to be the second field having the direct association relationship with the first field.

As can be seen from the above, for the field used in the historical search behavior, only the field in which the time difference between the time of performing the search using the field and the first time is smaller than the preset time difference and the frequency of the time difference smaller than the preset time difference in the historical search behavior of each user is greater than the preset frequency is used as the second field. Since the frequency of the user searching by using the first field and the second field together in a short time is high, that is, a large number of users frequently use the first field and the second field to search in a short time, the higher the association degree between the second field and the first field is theoretically determined, the higher the accuracy of the description information of the first field determined based on the second field with the high association degree is.

Further, after the above step a, the embodiment of the present invention further includes the following step E.

Step E: and taking the target frequency corresponding to the second field as the association relation weight between the first field and the second field.

Specifically, the weight indicates a degree of association between the first field and the second field, and the greater the target frequency, the higher the frequency of searching using both the first field and the second field in a short time, the higher the degree of association between the first field and the second field, and therefore the target frequency may be used as the weight of the first association relationship.

In one embodiment of the present invention, the step S103 may be implemented by the following step B and/or step C.

And B: and determining the entity with part or all of entity information being the same as the first field from the constructed knowledge graph as the first entity directly associated with the first field.

Specifically, the entity information may include an entity name, entity attribute information, and the like, and the entity attribute information may include an entity type, an entity generation time, a size of the entity information, and the like.

In an embodiment of the present invention, each entity may be traversed, and an entity whose part or all of the entity information is the same as the first field may be determined as the first entity, or an entity whose part or all of the entity information is a part of the first field may be determined as the first entity.

Specifically, the word segmentation processing may be performed on the first field, and the word segmentation processing result is compared with part or all of the entity information of the entity, so as to determine the first entity in the entity.

The word segmentation processing can be implemented by a word segmentation mode in the prior art, which is not limited in the embodiment of the present invention.

And C: and determining an entity with part or all of entity information having the same semantics with the first field from the constructed knowledge graph as the first entity directly associated with the first field.

In an embodiment of the present invention, semantic analysis may be performed on the first field, semantic analysis may be performed on all or part of the entity information of each entity, and the semantics of the part or all of the entity information of each entity may be compared with the semantics of the first field, so as to determine the first entity in the entities.

Specifically, the semantic analysis may be performed in a manner in the prior art, which is not limited in the embodiment of the present invention.

Referring to fig. 5, a flowchart of a second field description information obtaining method provided for the embodiment of the present invention is shown, and compared with the foregoing embodiment shown in fig. 1, the foregoing step S103 can be implemented by the following steps S103A-S103B.

S103A: and determining an identification sequence containing the identification of the first field according to the first association relation and a second association relation between the entities in the knowledge graph.

And the nodes represented by the adjacent identifiers in the identifier sequence directly have an association relationship, and the nodes are fields or entities.

Specifically, the identifier of the first field may be located at any position in the identifier sequence.

In addition, the identifier in the identifier sequence may be a number, a name, or the like of the node.

In one embodiment of the present invention, the step S103A can be implemented by the steps S103A1-S103A2, which will not be described in detail herein.

S103B: and determining fields except the first field and the second field corresponding to the identifiers in the identifier sequence as third fields, and determining entities except the first entity corresponding to the identifiers in the identifier sequence as second entities.

Specifically, since an association relationship directly exists between nodes corresponding to each pair of adjacent identifiers included in the identifier sequence, an association relationship exists between nodes corresponding to any two identifiers in the identifier sequence. And because the identification sequence comprises the identification of the first field, the nodes corresponding to the identifications in the identification sequence have a direct association relationship or an indirect association relationship with the first field. The third field and the second entity which have indirect association relation with the first field can be determined through the identification sequence.

Referring to fig. 6, a flowchart of a third method for obtaining field description information according to an embodiment of the present invention is shown, and compared with the foregoing embodiment shown in fig. 5, the above step S103A can be implemented by the following steps S103a1-S103a 2.

S103A 1: and selecting one node which directly has an association relation with the current target node as a new target node.

Specifically, in the case where the number of nodes as the past target nodes is smaller than the preset number, the above step S103a1 is repeatedly performed, and a new target node is repeatedly selected until the number of nodes as the past target nodes is not smaller than the preset number, and then step S103a2 is performed.

Wherein the initial value of the target node is the first field. That is, starting from the first field, a second field directly associated with the first field or a first entity directly associated with the first field is selected as a second target node. Further, a node directly associated with the second target node is selected as a third target node. And repeating the steps until a preset number of target nodes are obtained.

In an embodiment of the present invention, there may be a plurality of nodes directly associated with the target node, but there is only one node selected as the target node each time, and a new target node may be selected through any one of the following steps F to I.

Step F: and preferentially selecting one node which is not taken as the target node and has an incidence relation with the current target node as a new target node.

Step G: and preferentially selecting one node which is used as the target node and has an incidence relation with the current target node as a new target node.

Step H: and preferentially selecting a node which has the highest incidence relation weight with the current target node and directly has the incidence relation as a new target node.

Wherein, the association relation weight represents the association degree between the nodes.

Step I: and randomly selecting a node directly having an association relation with the current target node as a new target node.

In addition, the step F and the step H may be combined, that is, a node having the highest weight of the association relationship with the current target node among the nodes having the direct association relationship with the target node, which are not the target nodes, may be selected as the new target node.

And combining the step G and the step H, that is, selecting the node having the direct association relationship with the target node as the node having the highest weight of the association relationship with the current target node from the nodes passing through the target node as the new target node.

A new target node may also be selected in other manners, which is not described herein again in the embodiments of the present invention.

S103A 2: and generating an identification sequence which is arranged according to the first sequence and contains the identification of each target node.

Wherein the first order is: and taking each node as the sequence of the target node.

Specifically, each time step S103a1 is executed, a node directly associated with the current target node is selected as a new target node, and therefore, an association relationship directly exists between nodes selected as target nodes twice before and after. Therefore, in the identifier sequence generated according to the first sequence, the nodes corresponding to the adjacent identifiers directly have an association relationship, and the nodes corresponding to the non-adjacent identifiers can directly or indirectly have an association relationship. In addition, since the initial value of the target node is the first field, the first field serves as a target node, and the generated identification sequence includes the first identification of the first field, that is, each node corresponding to each identification in the identification sequence except the first field has a direct or indirect association relationship with the first field.

In addition, the condition for ending the repeated execution of the process of step S103a1 described above is that the number of nodes as the target passing nodes is not less than the preset number, and therefore the number of nodes as the target passing nodes is accumulated up to the preset number. That is, the identifier sequence formed by the identifiers of the target nodes contains a preset number of identifiers.

As can be seen from the above, the first field is used as an initial target node, and other nodes are sequentially determined as new target nodes according to the association relationship between the nodes, and each determined new target node has a direct or indirect association relationship with the first field. The identifiers of the nodes are arranged according to the sequence as the target node to form an identifier sequence, and the nodes corresponding to the identifier sequence have direct or indirect association with the first field. Therefore, the identification sequence obtained by the embodiment of the invention can represent the entity or the field which has the association relation with the first field.

In another embodiment of the present invention, the initial value of the target node may be other nodes, and on the basis of the above steps S103a1-S103a2, the identification sequence may be generated, and the identification sequence including the first identification may be selected from the generated identification sequences.

Corresponding to the foregoing field description information obtaining method, referring to fig. 7, a schematic structural diagram of a first field description information obtaining apparatus according to an embodiment of the present invention is shown, where the apparatus includes: .

A field extraction module 701, configured to obtain a user search behavior for video content, and extract a first field for searching from the user search behavior;

a field determining module 702, configured to determine, based on historical search behavior for video content, a second field directly associated with a first field, where a time difference between a first search time and a second search time is less than or equal to a preset time difference, where the first search time is: the time for searching by using the first field in the historical searching action is as follows: time to search using the second field in the historical search behavior;

an entity searching module 703, configured to search for a first entity directly associated with the first field from a constructed knowledge graph, where the knowledge graph is: composed of entities associated with video content;

an entity field searching module 704, configured to search, according to the first entity, a second entity indirectly associated with the first field from the knowledge graph, and determine, based on a first association between the second field and a known field, a third field indirectly associated with the first field;

an information obtaining module 705, configured to obtain description information of the first field according to target description information, where the target description information includes: the description information of the second field, the description information of the third field, the description information of the first entity, and the description information of the second entity.

In an embodiment of the present invention, the entity searching module 703 is specifically configured to:

determining part or all of entities with the same entity information as the first field from the constructed knowledge graph as first entities directly associated with the first field;

and/or

And determining an entity with part or all of entity information having the same semantic meaning as the first field from the constructed knowledge graph as the first entity directly associated with the first field.

Referring to fig. 8, a schematic structural diagram of a second field description information obtaining apparatus according to an embodiment of the present invention is shown, and compared with the embodiment shown in fig. 7, the entity field searching module 704 includes:

a sequence determining submodule 704A, configured to determine, according to a first association relationship and a second association relationship between entities in the knowledge graph, an identification sequence including an identifier of the first field, where an association relationship directly exists between nodes represented by adjacent identifiers in the identification sequence, and the nodes are fields or entities;

and the entity field determining submodule 704B is configured to determine, as a third field, a field except the first field and the second field corresponding to the identifier in the identifier sequence, and determine, as a second entity, an entity except the first entity corresponding to the identifier in the identifier sequence.

As can be seen from the above, since the association relationship directly exists between the nodes corresponding to each pair of adjacent identifiers included in the identifier sequence, the association relationship exists between the nodes corresponding to any two identifiers in the identifier sequence. And because the identification sequence comprises the identification of the first field, the nodes corresponding to the identifications in the identification sequence have a direct association relationship or an indirect association relationship with the first field. The third field and the second entity which have indirect association relation with the first field can be determined through the identification sequence.

Referring to fig. 9, a schematic structural diagram of a third field description information obtaining apparatus provided in the embodiment of the present invention, compared with the foregoing embodiment shown in fig. 8, the sequence determining sub-module 704A includes:

a node selecting unit 704a1, configured to repeatedly perform, according to a first association relationship and a second association relationship between entities in the knowledge graph, a step of selecting, as a new target node, a node having an association relationship directly with a current target node until the number of nodes serving as target nodes is not less than a preset number, where an initial value of the target node is the first field;

a sequence generating unit 704a2, configured to generate an identification sequence including identifications of target nodes arranged in a first order, where the first order is: and taking each node as the sequence of the target node.

In an embodiment of the present invention, the node selecting unit 704a1 is specifically configured to:

repeatedly selecting a new target node in any one of the following modes according to the first incidence relation and a second incidence relation between the entities in the knowledge graph until the number of the nodes serving as the target nodes is not less than a preset number:

preferentially selecting a node which is not used as a target node and has an incidence relation with the current target node directly as a new target node;

preferentially selecting a node which is used as a target node and has an incidence relation with the current target node directly as a new target node;

and preferentially selecting a node which has the highest association relation weight with the current target node and directly has the association relation as a new target node, wherein the association relation weight represents the association degree between the nodes.

In an embodiment of the present invention, the field determining module 702 is specifically configured to:

determining a field with a target frequency greater than a preset frequency as a second field based on historical search behavior aiming at video content, wherein the target frequency is as follows: the time difference between the time of searching by using the field in the historical searching action and the first searching time is less than the frequency of the preset time difference.

In one embodiment of the present invention, the apparatus further comprises:

and the weight determining module is used for taking the target frequency corresponding to the second field as the association relation weight between the first field and the second field.

In one embodiment of the invention, the knowledge graph is constructed by the following graph construction modules;

the map building module is used for:

acquiring entities related to video content, and determining the association relationship among the entities;

generating a triple based on the association relationship among the entities, wherein the elements in the triple are respectively: entity identifications of two entities with incidence relation and the type of the incidence relation between the entities;

constructing the knowledge-graph based on the generated triples.

The embodiment of the present invention further provides an electronic device, as shown in fig. 10, which includes a processor 1001, a communication interface 1002, a memory 1003 and a communication bus 1004, wherein the processor 1001, the communication interface 1002 and the memory 1003 complete mutual communication through the communication bus 1004,

a memory 1003 for storing a computer program;

the processor 1001 is configured to implement the method steps of any of the above field description information obtaining methods when executing the program stored in the memory 1003.

When the electronic device provided by the embodiment of the invention is applied to obtain the description information of the first field, the scheme provided by the embodiment of the invention determines the first entity in the knowledge graph, which directly has an association relationship with the first field, in addition to the second field which directly has an association relationship with the first field based on the historical search behavior, and also determines the third field and the second entity which indirectly have an association relationship with the first field. Therefore, by introducing the constructed knowledge graph, the first field not only has an association relationship with the known field, but also has an association relationship with the entity, and the association relationship with the first field is rich. Therefore, the nodes such as the fields and the entities having the association relation with the first field are abundant, and the description information of the first field obtained according to the description information of the abundant nodes is accurate.

The communication bus mentioned in the above terminal may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.

The communication interface is used for communication between the terminal and other equipment.

The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.

The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.

In another embodiment of the present invention, a computer-readable storage medium is further provided, in which a computer program is stored, and the computer program, when executed by a processor, implements the method steps of any of the above field description information obtaining methods in the above embodiments.

When the computer program stored in the computer-readable storage medium provided by this embodiment is applied to obtain the description information of the first field, the solution provided by this embodiment of the present invention determines, in addition to the second field that directly has an association relationship with the first field based on the historical search behavior, the first entity that directly has an association relationship with the first field in the knowledge graph, and also determines the third field that indirectly has an association relationship with the first field and the second entity. Therefore, by introducing the constructed knowledge graph, the first field not only has an association relationship with the known field, but also has an association relationship with the entity, and the association relationship with the first field is rich. Therefore, the nodes such as the fields and the entities having the association relation with the first field are abundant, and the description information of the first field obtained according to the description information of the abundant nodes is accurate.

In a further embodiment of the present invention, there is also provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the method steps of any of the above field description information obtaining methods of the above embodiments.

When the computer program product provided by the embodiment is executed to obtain the description information of the first field, the scheme provided by the embodiment of the invention determines the first entity in the knowledge graph, which directly has an association relationship with the first field, in addition to determining the second field which directly has an association relationship with the first field based on the historical search behavior, and also determines the third field which indirectly has an association relationship with the first field and the second entity. Therefore, by introducing the constructed knowledge graph, the first field not only has an association relationship with the known field, but also has an association relationship with the entity, and the association relationship with the first field is rich. Therefore, the nodes such as the fields and the entities having the association relation with the first field are abundant, and the description information of the first field obtained according to the description information of the abundant nodes is accurate.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, as for the apparatus, the electronic device, the computer-readable storage medium and the computer program product, since they are substantially similar to the method embodiments, the description is relatively simple, and in relation to them, reference may be made to the partial description of the method embodiments.

The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims

1. A field description information obtaining method, characterized in that the method comprises:

2. The method of claim 1, wherein the searching the constructed knowledge-graph for the first entity directly associated with the first field comprises:

and/or

3. The method of claim 1, wherein the searching, according to the first entity, for a second entity indirectly associated with the first field from the knowledge-graph, and determining a third field indirectly associated with the first field based on the second field and an association between known fields comprises:

determining an identification sequence containing the identification of the first field according to a first incidence relation and a second incidence relation between entities in the knowledge graph, wherein incidence relations directly exist between nodes represented by adjacent identifications in the identification sequence, and the nodes are fields or entities;

and determining fields except the first field and the second field corresponding to the identifiers in the identification sequence as third fields, and determining entities except the first entity corresponding to the identifiers in the identification sequence as second entities.

4. The method of claim 3, wherein determining an identification sequence containing an identification of the first field comprises:

repeatedly executing the step of selecting one node directly having an association relation with the current target node as a new target node until the number of the nodes serving as the target nodes is not less than the preset number, wherein the initial value of the target node is the first field;

generating an identification sequence which is arranged according to a first sequence and contains the identification of each target node, wherein the first sequence is as follows: and taking each node as the sequence of the target node.

5. The method according to claim 4, wherein selecting a node having an association relationship directly with the current target node as the new target node comprises:

selecting a new target node by any one of the following means:

6. The method of any of claims 1-5, wherein determining the second field directly associated with the first field based on historical search behavior for video content comprises:

7. The method of claim 6, wherein after determining, as the second field, a field with a target frequency greater than a preset frequency based on historical search behavior for video content, the method further comprises:

and taking the target frequency corresponding to the second field as the association relation weight between the first field and the second field.

8. The method according to any one of claims 1-5, wherein the knowledge-graph is constructed by;

constructing the knowledge-graph based on the generated triples.

9. A field description information obtaining apparatus, characterized in that the apparatus comprises:

10. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;

a memory for storing a computer program;

a processor for implementing the method steps of any of claims 1 to 8 when executing a program stored in the memory.